C2KD: Cross-Lingual Cross-Modal Knowledge Distillation for Multilingual Text-Video Retrieval
Lists
Only you see this
Abstract
Multilingual text-video retrieval methods have improved significantly in recent years, but the performance for other languages lags behind English. We propose a Cross-Lingual Cross-Modal Knowledge Distillation method to improve multilingual text-video retrieval. Inspired by the fact that English text-video retrieval outperforms other languages, we train a student model using input text in different languages to match the cross-modal predictions from teacher models using input text in English. We propose a cross entropy based objective which forces the distribution over the student's text-video similarity scores to be similar to those of the teacher models. We introduce a new multilingual video dataset, Multi-YouCook2, by translating the English captions in the YouCook2 video dataset to 8 other languages. Our method improves multilingual text-video retrieval performance on Multi-YouCook2 and several other datasets such as Multi-MSRVTT and VATEX. We also conducted an analysis on the effectiveness of different multilingual text models as teachers.
BibTeX
@inproceedings{rouditchenko2023c,
author = {Andrew Rouditchenko and Yung-Sung Chuang and Nina Shvetsova and Samuel Thomas and Rogério Feris and Brian Kingsbury and Leonid Karlinsky and David Harwath and Hilde Kuehne and James R. Glass},
title = {{C2KD: Cross-Lingual Cross-Modal Knowledge Distillation for Multilingual Text-Video Retrieval}},
booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
year = {2023},
doi = {10.1109/icassp49357.2023.10094821},
}